Python 3 adoption

http://caniusepython3.com 's count is misleading because it appears to include almost all packages hosted in PyPI. Many of these haven't been updated in many years and many were never very serious projects. https://python3wos.appspot.com is much more fair, cataloging 200 widely-used Python libraries. Only 17 of these don't have a Python 3 version yet. I remember when 80% of that wall was red.

I use Arch. Arch shipped Python 3 as the default Python in 2011 iirc. It was bumpy for a while, but now I scarcely ever have problems with it. I think the most important step for Python 3's adoption is to get RHEL and Debian (and their respective downstreams) to promote Python 3 to the default. It's time; the old arguments that it's an inordinate amount of work to support Py3 as the default don't really hold anymore. RHEL and Debian have huge cultural influence by deeming a version "enterprise-ready" and making it default.

Here's to Python 3. We're on the cusp of killing off 2 for new development. Let's not backslide.

At this point it may require more leadership than an informational PEP. They let the community take their time adopting it, and it still doesn't feel like there is any rush to adoption. And that create a split in the community, fractured module dependencies, and 5 years later it still doesn't seem to have found majority support. And for what? Most users probably won't notice more than Unicode strings, a few syntax changes such as the print function, maybe the exception handling. It didn't change any of the language fundamentals, or address any common criticism of the language such as speed, concurrency or the type system. There are nice changes in Python 3 under the hood, but I'm starting to wonder if the cost was truly worth it.

The Python community beeing so split and "backward-oriented" is really the main reason why I wouldn't consider Python, though I like many aspects of the language.